The general resolution to abolish non-free is flawed in a number of ways.
It weakens the social contract
==============================
The social contract is one of the foundations on which Debian users
base their expectations for Debian's future directions. Up until now,
Debian users could rely on having both a fully functional and free
distribution available to them and a number of non-free components that
they may require. While we will, with the proposed changes, continue to
acknowledge our users' needs for non-free software, we will no longer
make any attempt to satisfy them.
Futhermore, the success of this proposal would set a precedent that
the Debian developers may, by a simple majority vote, modify the social
contract as they wish, without making any attempt to determine if this
is truly in the interests of their users, or even the free software
community at large.
It offers no tangible benefits
==============================
This proposal is purely negative: it seeks to remove all the functionality
that the non-free component offers Debian users, without offering anything
that will directly improve a user's experience of the Debian distribution,
or make participation in the Debian project easier or more productive.
It creates both immediate and long-term tangible drawbacks
==========================================================
The initial result of this proposal will be to cease updating non-free
software which will make maintenance and installation of contrib packages
significantly more difficult, since there will be no reliable site for
the non-free software many contrib packages depend upon. In the longer
term, it will make it impossible to assure users that contrib software
can be installed and is usable.
Similarly, it will make it impossible for Debian to assure its users that
particular items of non-free software will be available. It is commonly
held that this change will make non-free software for Debian no harder
to find, but this is a best case scenario assuming a single archive of
comparable quality to Debian's is set up. If multiple archives are setup,
or the archives are of lower quality (eg they do not follow a technical
policy document, eg, or they do not check uploads for authenticity),
using non-free software on a Debian system will become significantly
more difficult.
It's premature
==============
non-free software is still necessary: there is still free software based
on it (as evidenced by contrib), and many of our users still require it.
The premise of section five of the social contract still holds: some
of our users do still require the use of programs that don't conform to
the Debian Free Software Guidelines. Which programs are required and have
no free replacements has changed and continues to change, but the fact
that some programs are still required has not.
It forks the Debian infrastructure
==================================
The Debian infrastructure (the mirror network, the BTS, the archive,
the keyring, the new-maintainer procedure, technical policy, decision
making policy) is well established and instrumental in making Debian
as maintainable as it is, and will be necessary to maintain non-free
software at a similar level.
But since Debian will no longer distribute non-free software, much of
this infrastructure will become unavailable: the mirror network and the
archive directly, the BTS will require modifications and extra ongoing
maintenance to be useful for packages Debian refuses to distribute,
a seperate new-maintainer system and decision making hierarchy will
be needed for maintainers that do not wish to package free software as
well as non-free software, concerns raised by non-free packages (such
as naming conflicts) will tend not be addressed by Debian and must thus
be worked around instead of fixed.
As such, those maintainers who wish to continue to support non-free
software will need to duplicate many of these items, either incompletely,
leading to more difficult maintenace and a poorer user experience,
or completely, leading to significant duplicated and wasted effort.
Further to this, many of the rationales proposed for this change are
also flawed on a number of matters:
Confusion between the Debian project and the Debian distribution
================================================================
It is undeniably true that the Debian GNU/Linux and GNU/Hurd distributions
solely contain free software. This does not, however, imply that the
Debian project is exclusively limited to distributing free software,
and indeed the Debian project has distributed non-free software and has
codified that practice in its social contract. Distributing non-free
software is not evidence of hypocrisy or compromised ideals, it is a
corollary of the needs of our distribution's users, and the project's
support for those users.
Confusion between the Debian project and Debian developers
==========================================================
The social contract does not dictate what individual developers must
work on, only what the project as a whole will try to achieve. Just
as individual developers are not required to only ever write free
software, individual developers are similarly not required to work on
non-free software or assist with its infrastructure. With or without
the commitment to non-free software contained in the social contract,
developers may completely ignore non-free software, without decreasing
their contribution to the distribution or the project in any way.
The effects of this proposal on the free software community
===========================================================
Debian, and Debian developers, have long been advocates of free software,
by writing new free software, by improving existing free software,
and by encouraging existing non-free software to change its license
terms to a free alternative. Some supporters of this proposal see it
as a continuation of this latter trend in that it will make authoring
non-free software less attractive if it is not distributed by Debian.
They neglect to mention that Debian is already highly successful at
encouraging non-free software authors to relicense, that distributing
software as non-free rather than part of the distribution itself is
already a successful disincentive, that often it is the maintainer of the
non-free package who initiates the discussion even after it's packaged
for non-free, and that seeing that a package is non-free often in and
of itself encourages users to look for a free replacement.
Respectfully submitted,
aj
--
Anthony Towns <[EMAIL PROTECTED]> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG encrypted mail preferred.
``We reject: kings, presidents, and voting.
We believe in: rough consensus and working code.''
-- Dave Clark
PGP signature
